Cathode Ray Typesetting - Just a quick question

dsgoen's picture

Does anybody know the approximate lpi scan resolution of CR typesetting circa 1975? I'm trying to get a feel for how it would translate to current dpi-rated devices.


Don McCahill's picture

I don't know the answer, but something in my brain tells me that the fonts of the time were built on something like a 13x9 or 14x9 grid. I remember having an old Hercules card that let you edit the fonts, one pixel at a time. Great way to waste a few hours.

Theunis de Jong's picture old Hercules card...

In 1975?

I started typesetting on an AM VariTyper -- some 10 years later -- and actually never knew its "resolution". I *do* know a few tricks you could do with it: overflash, to make type appear bolder, and super-sizing, right until the edges grew blurry and the outer parts were chopped off.

Ah now I remember -- that was one machine later. For its precursor (and the very first thing I had ta do in the studio), I had to manually change the wheel containing (film) negatives of fonts [*]. No resolution needed, it scaled fonts by moving the wheel back and forth.

[*] One font per wheel. First, insert disk labeled "Roman" and image a page (or rather, a strip of type). Remove disk, insert "Italic", re-program computer and image italic text. Repeat with other required fonts until bored to death or making a mistake.

cuttlefish's picture

y'know, these are just the sort of things that need to be mentioned in the 20th century typography developments thread.

Short lived though these technologies were, they were an important step in changing what is thought of as type, from individual pieces of wood and metal to the mathematical outlines today, the intermediate steps transitioning from physical objects to representations of shapes happened here.

paragraph's picture

Don: your Hercules card was in a PC computer, much later. Theunis worked on a optical projection machine. The machines in question would be something like Compugraphic 4961 Phototypesetting Machine. Alas I have no idea what the scan resolution correspods to. Will? Charles?

cuttlefish's picture

Maybe something useful here:
I didn't spot exactly something mentioning the scan resolution of the video system, but I might have missed it.

charles ellertson's picture

The Linotron 202 was about 960 lines per inch. Early imagesetters for PostScript fonts were frequently used at 1200 dpi, given our experience with the 202, we ran our Birmy imagesetter at 1,000 dpi (for type only).

The 2,400 and 3,600 dpi output you see today are only needed because images and type are in the same file. And worse, you can't "increase the exposure" today, because that would translate into greater dotgain for the images. With the older techonologies, like the 202, we'd increase the exposure when setting a font with fine lines (like Perpetua), and decrease it with a front that didn't have fine lines. The overall weight of the font depended on things other than it's design, namely the exposure when output, the development used for the repro paper, the exposure when the negative was made, less on the development of the negative, but again, the exposure when the plate was made.

Sort of hard to compare. . .

charles ellertson's picture


Sort of off-topic, but I too began computer comp on an AM-Varityper -- the CompSet. Our machine used disks that (standard) had four fonts on them. But we also had AM make us up custom font disks, all ligatures, all the small caps, etc. on one disk. I'd make them redraw a few characters to boot. That took us down to roman & italic on one font. It was also a bit pricy; we had custom disks for Times Roman and Sabon. Plus later, a pretty good Galliard. These were Monotype masters, and both the Times and Sabon were quite nice.

The ComSet had a one-line buffer. Once you set a line of type, it was gone. You could see approximately the next line. If, one line yet later in a paragraph, you came on a line ending with an unbreakable word ("strength" "women" etc.), you were going to run the chapter out again. In fact, that was the normal way of working. You set and ran out a chapter, then marked up the repro so that when you set it the next time, you had accounted for all the problems. You absolutely had to guess right about how things were going to fit, or you had to run it out a third time. Repro paper wasn't cheap; errors were costly.

When the compEdit came along, it had a buffer. You could set a whole paragraph before sending it off. Thought I'd gone to heaven. It also had "reverse feed", which made setting display equations much easier. I use to set them using tabs, and that was a bitch when a line was gone as soon as you set it.

The plus to the Varityper equipment was (1) it didn't break so much, and (2) the measure stayed pretty constant. I'm not sure "measure" was a recognizable term for the VIP -- or the 202, for that matter. When we ran corrections on a job with a 202, we'd run off a couple test lines, so as to keep the "measure" to within 3 points of the original. Ah, the wanderings of horizontal amplifiers . . .

The early days of photocomp weren't much different than metal in terms of operator skill needed for good quality work. But there were no heavy trays of metal, probably the most important thing in getting women into the composing room. Health issues changed too, no lead, but I know a number of comps who spent long days at those early terminals who now have lymphoma.

But all this is off the topic . . .

johndberry's picture

Charles –

There was definitely a difference in output quality on a Linotronic between 1200dpi and 2400dpi, at least if you were using ITC Garamond (an unfortunately common text face in those days). That wasn't the only typeface where you could see a difference, but it was the one I used as a benchmark, because it was so obvious. At 1200dpi, the strokes just looked anemic.

You remind me of when I would send files to a service bureau for output, and for certain typefaces I would ask them to boost the density. One shop was great for images but not so good for text; another gave me good strong density for text-only output, which was most of what I was doing. I would never have used digital Centaur for one book if I hadn't been getting it output to RC paper and had the option of asking for higher density; otherwise, the digital font isn't strong enough.

But the Linotron 202 wasn't introduced until 1978, so that doesn't really answer the original question.


charles ellertson's picture

I don't know if it is available online, but Dick Angstadt (Keystone Typesetting) and I had a longish discussion on this in The Glossary of Typesetting Terms under *Output Resolution*. I'm not going to type in the whole page+, but after I'd talked about adjusting the exposure with the 202, Dick said "Many imagesetter fonts have been designed to perfectly match metal type fonts without compensating for the slight spread of the ink when the characters were printed letterpress. At very high resolution, these fonts appear too thin when printed by lithographic process. Outputting at a lower resolution (about 1,200 dpi) helps these fonts regain some of their original character."

I agreed, and added "A similar effect [gaining density through exposure] can be achieved with imagesetters. For example, when Monotype Bembo was released in PostScript form, it appeared too thin at small sizes. We experimented with different output resolutions -- 1,200 dpi, 900 dpi, and 600 dpi -- on a Linotron 200 imagesetter and with different exposures at each resolution. we found that increasing the exposure at a 1,200 setting produced a more usable image than simply lowering the resolution setting . . ."

So yes, if you are trying to reproduce *the font as drawn*, higher resolution is better. But if the drawing is wrong (as so many are), it's the wrong way to go after a certain resolution is reached -- about 1,200 dpi, in my testing. BTW, when we went to a DTP workflow, even a font like Scala needed an increase of 2/1000 in character weight to match what was printed with repro-negative-plate workflows. And of course, if Majoor was paying attention to what was *printed* when he drew the font, he designed Scala when repro-negative-plate was the dominant workflow.

* * *

Our 202 had a plate inside the doors which gave a date of sometime in the 1960s, but maybe that was just for the computer part. At any rate, the 505 and 202 (I never knew much about the 303) were the first photocomp machines I knew of that used a video tube to expose the paper. The ones before these that I'm familiar with used a projection system. Not to say there wasn't one.

dsgoen's picture

Thank you for all of your help.Cuttlefish is right, we need to document all of these obsolete early digital systems. They changed the world of typesetting, but they are rapidly become lost knowledge.

paragraph's picture

The original question relates to the era just before PostScript and desktop type and layout, so the Linotronics are unfortunately out. If I still remember correctly, in their case the 1200 ppi was deemed fine for line bromides or film, and the "high" resolution of 2400+ ppi was seen as necessary for half-tone screens to achieve the 256 levels of gray at 150 lpi.

There were pre-PS CRT Linotype machines, however. Apart from the already mentioned Compugraphic there are:

1974 - ... the cathode ray typesetting machines "Linotron 303/TC" and “Linotron 505/TC”

Still no notion of the resolution though.

raph's picture

I seem to recall 1000 dpi or so being most common, and a snippet from Typesetting and Composition seems to confirm it - the Linotron 202 was 975 lines per inch in standard resolution mode, and 1950 in high-resolution. That was launched in 1978, not the 1974 you requested, but I don't believe there were huge changes that decade.

Books are actually a pretty good source for this kind of archival or historic information, even more so now that they're commonly online :)

bowerbird's picture

wow. nostalgia rush.

when i rolled my first page out of a linotronic,
i really felt like i had crossed a big threshold...

i can still recall the smell of it.


kentlew's picture

> Cuttlefish is right, we need to document all of these obsolete early digital systems.

FWIW, some of this technology is preserved at the Museum of Printing in North Andover, MA. If you don't find your answers online, you might contact someone at the museum to see if they know the answer. Unfortunately, Gardner LePoer retired recently as director, due to health issues, and I don't know if he's been replaced yet.

paragraph's picture

Charles: At any rate, the 505 and 202 (I never knew much about the 303)

Sorry, I missed that bit, my reading was superficial again. You should be able to figure the resolution out, please?

charles ellertson's picture

As I said earlier, we determined (empirically) the standard resolution of our 202 to be 960 lpi. Raph posted that it was specified as 975 lpi, so that sounds like a good number.

Best I can remember from gossiping from the Linotype repairman -- we saw a lot of him -- the 505 was a custom-built machine, for high volume work. According to the repairman, Reader's Digest had a 505 in the old days when they set both books & the magazine in house. I doubt the resolution was any higher, but who knows. I was interested in Reader's Digest machine because the magazine was set in Granjon, & their Granjon looked just a bit better than the 202 font. But it was probably only the paper used . . . again, it was just so long ago.

Syndicate content Syndicate content